Maridebart cafraglutide — Amgen calls it MariTide — is the first drug in the weight-loss race built on an antibody rather than a peptide, and it is being tested as a once-a-month injection. This page reads what has been published and registered for it under all three of its names (maridebart cafraglutide, MariTide, AMG 133), as of 28 September 2026. It is not approved anywhere.
For the scientific puzzle it raises — why blocking the GIP receptor and activating it both seem to help weight loss — see GIP and GLP-1. For the other drugs in late-stage trials, see CagriSema, amycretin and retatrutide.
What it is
Approved GLP-1 medicines are peptides, small protein chains. Semaglutide weighs about 4 kDa and is injected weekly; the what-is-a-peptide page sets out the class's size range.
MariTide is built differently:
- The backbone is a monoclonal antibody that binds and blocks the GIP receptor.
- Two synthetic GLP-1 agonist peptides are chemically attached to it — conjugated, not genetically fused, which Amgen's discovery paper says lets chemists modify the peptides freely (J Med Chem 2026).
- The antibody carries the peptides for weeks. Antibodies are recycled in the body by the neonatal Fc receptor and are too large for the kidneys to filter, which is why the design supports monthly dosing.
| MariTide | Tirzepatide (Zepbound) | Semaglutide (Wegovy) | |
|---|---|---|---|
| Type | antibody–peptide conjugate | peptide | peptide |
| GLP-1 receptor | activates | activates | activates |
| GIP receptor | blocks | activates | — |
| Schedule tested | every 4 or 8 weeks | weekly | weekly (and daily tablet) |
| Status (Sept 2026) | phase 3 | approved | approved |
The phase 2 trial
The published phase 2 (NCT05669599; NEJM 2025, Amgen-funded) enrolled 592 people in two cohorts and ran 52 weeks. Doses were 140, 280 or 420 mg every four weeks, or 420 mg every eight weeks, with some arms starting straight at the full dose and others escalating over 4 or 12 weeks.
| Cohort | People | Weight change on MariTide (range across arms) | Placebo |
|---|---|---|---|
| Obesity | 465 (63% female, mean BMI 37.9) | −12.3% to −16.2% | −2.5% |
| Obesity + type 2 diabetes | 127 (42% female, mean BMI 36.5) | −8.4% to −12.3% | −1.7% |
These are treatment-policy figures, which count everyone randomised whether or not they kept taking the drug. Higher figures circulating online — around 17% to 20% — come from other estimands (people who stayed on treatment) or from phase 1, and are not the same measurement. In the diabetes cohort, HbA1c fell 1.2 to 1.6 points against a 0.1-point rise on placebo.
Two features of that result are worth reading carefully:
- The diabetes gap. People with type 2 diabetes lost roughly a third less weight — the same pattern seen with every GLP-1 medicine, set out on GLP-1 before and after.
- Every-eight-week dosing was tested in only one arm (420 mg), so whether a two-monthly schedule holds up is a phase 3 question. How weight loss levels off across the class is on GLP-1 plateau.
Side effects. The abstract reports gastrointestinal adverse events as common, and less frequent with dose escalation and a lower starting dose; no unexpected safety signals. The per-arm rates are in the full paper and are not reproduced here.
The phase 3 programme: MARITIME
On 28 September 2026 ClinicalTrials.gov held 28 records for the drug: 15 phase 1, 3 phase 2, 10 phase 3 — none with results posted. The phase 3 set:
| Trial | People | Population | Primary completion |
|---|---|---|---|
| MARITIME-1 (NCT06858839) | 3,853 | obesity, no diabetes | 21 Jan 2027 |
| MARITIME-2 (NCT06858878) | 1,105 | obesity + type 2 diabetes | 21 Jan 2027 |
| MARITIME-3-J (NCT06987695) | 279 | Japan | 27 Apr 2027 |
| MARITIME-OSA-1 (NCT07225686) | 250 | sleep apnoea, on PAP | 3 Sep 2027 |
| MARITIME-OSA-2 (NCT07226765) | 250 | sleep apnoea, not on PAP | 29 Sep 2027 |
| Switching trial (NCT07575399) | 300 | switching from another GLP-1 | 3 Jan 2028 |
| Extension trials (NCT07684235, NCT07684144) | 3,200 / 950 | long-term follow-on | 26 Dec 2027 |
| MARITIME-HF (NCT07037459) | 5,056 | heart failure (preserved or mildly reduced ejection fraction) | 30 Jun 2028 |
| MARITIME-CV (NCT07037433) | 12,800 | cardiovascular outcomes | 30 Jun 2028 |
Three things stand out in that register:
- No registered head-to-head against tirzepatide or semaglutide. Every efficacy trial above is placebo-controlled; the switching trial moves people from another GLP-1 but is not a comparison of weight loss. So even after 2027, how MariTide compares with the approved drugs will be a cross-trial comparison, with the caveats on best GLP-1 for weight loss.
- The sleep-apnoea trials copy Zepbound's route to its second indication (see GLP-1 and sleep apnoea), including the on-PAP / not-on-PAP split.
- The register grew by two records in three days. On 25 September our GIP page counted 26 studies; the two long-term extension trials, which started enrolling in July 2026, account for the difference. Registry counts move, which is why every count here carries its date.
What is not known
- Bone. The GIP receptor sits on bone-forming and bone-resorbing cells; mice lacking it have weaker cortical bone, and people carrying a loss-of-function GIPR variant have been linked to lower bone density. A 2026 review of the drug lists bone mineral density as a side effect that "need[s] to be discussed" (Antib Ther 2026). No published trial reports it. For comparison, the current Wegovy label records more hip and pelvis fractures in women on Wegovy than on placebo in its cardiovascular trial (1% against 0.2%).
- Lean mass. No published MariTide trial reports body composition.
- The GIP question itself. Why an antagonist and an agonist at the same receptor both add to GLP-1 weight loss has no settled explanation.
- Dose, price and timing. No filing has been announced. Per-compound dosing is not covered on this site.
This page reports trial records and does not recommend any medicine. What is approved today, and what it costs, is on the GLP-1 drug list and GLP-1 cost.
Sources
- Jastreboff AM et al. Once-monthly maridebart cafraglutide for the treatment of obesity — a phase 2 trial. N Engl J Med 2025;393(9):843–857. doi:10.1056/NEJMoa2504214 (PMID 40549887; correspondence PMIDs 41337722–41337724).
- Discovery of AMG 133, a GIPR antagonist and GLP-1R agonist antibody-drug conjugate for the treatment of obesity. J Med Chem 2026 (PMID 41941715).
- Design and therapeutic rationale of antibody–peptide conjugates: insights from maridebart cafraglutide (AMG133). Antib Ther 2026 (PMID 42592044, open access PMC13463644) — mechanism, bone-density discussion, phase 1 summary.
- ClinicalTrials.gov API v2, intervention "maridebart OR AMG 133 OR MariTide", queried 28 September 2026: 28 records; NCT05669599, NCT06858839, NCT06858878, NCT06987695, NCT07225686, NCT07226765, NCT07575399, NCT07684235, NCT07684144, NCT07037459, NCT07037433.